Fixing `name 'torch' is not defined` in `bitsandbytes` integration (#31243)

Fixed torch definition error

Fixing `name 'torch' is not defined` in `bitsandbytes` integration (#31243)
Fixed torch definition error
63fb253d · James Braza · GitHub · 66875ac0 · 63fb253d
Unverified Commit 63fb253d authored Jun 04, 2024 by James Braza Committed by GitHub Jun 05, 2024
Show whitespace changes
Inline Side-by-side

Showing with 1 addition and 1 deletion

src/transformers/integrations/bitsandbytes.py src/transformers/integrations/bitsandbytes.py +1 -1

No files found.
--- a/src/transformers/integrations/bitsandbytes.py
+++ b/src/transformers/integrations/bitsandbytes.py
@@ -332,7 +332,7 @@ def get_keys_to_not_convert(model):


 # Copied from PEFT: https://github.com/huggingface/peft/blob/47b3712898539569c02ec5b3ed4a6c36811331a1/src/peft/utils/integrations.py#L41
-def dequantize_bnb_weight(weight: torch.nn.Parameter, state=None):
+def dequantize_bnb_weight(weight: "torch.nn.Parameter", state=None):
    """
    Helper function to dequantize 4bit or 8bit bnb weights.